DDP accelorator should have 'trainer=none' as default

[edenlightning]

Add to DDP accelerator 'trainer=None' as a first argument (otherwise I cannot pass it to the Trainer instantiation)

https://github.com/PyTorchLightning/pytorch-lightning/blob/master/pytorch_lightning/accelerators/ddp_accelerator.py#L49

class DDPAccelerator(Accelerator):
    def __init__(self, trainer, cluster_environment=None, ddp_plugin=None)